ClawBio and BioClaw
These are complementary tools within the OpenClaw bioinformatics ecosystem: ClawBio provides the foundational skill library for bioinformatics agents, while BioClaw builds a higher-level research assistant interface that leverages those skills.
About ClawBio
ClawBio/ClawBio
🦖 ClawBio — The first bioinformatics-native AI agent skill library. Local-first. Reproducible. Built on OpenClaw.
Provides 39 pre-built skills that wrap genomic databases (UK Biobank, PharmGKB, PGS Catalog, GWAS databases) and bioinformatics pipelines (Kraken2, HUMAnN3, ancestry PCA) as reproducible agent tasks, eliminating hallucination in pharmacogenomics, variant lookup, and polygenic risk scoring. Built on OpenClaw's agent framework with Python 3.10+, each skill auto-exports a reproducibility bundle (`commands.sh`, `environment.yml`, SHA-256 checksums) so results can be re-run offline without the agent. Integrates Galaxy tool ecosystem (~8,000 tools) and ships with a CC0-licensed reference genome (Corpasome: 23andMe SNP chip + 30x whole-genome VCF) for deterministic local-first execution on user machines.
About BioClaw
Runchuan-BU/BioClaw
AI-Powered Bioinformatics Research Assistant. Built on OpenClaw.
Leverages Claude's tool-use capabilities to integrate bioinformatics workflows—BLAST, FastQC, PyMOL visualization, and PubMed search—directly into WhatsApp, Slack, and Feishu group chats via a Docker-containerized agent. Built on the NanoClaw architecture with pluggable skill modules (developed in a separate BioClaw_Skills_Hub for staging before promotion to core). Supports multi-provider LLM backends including Anthropic, OpenRouter, and OpenAI-compatible endpoints, with per-thread working directories and workflow shortcuts for persistent research sessions.
Scores updated daily from GitHub, PyPI, and npm data. How scores work